ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/BBClone/branches/0.6.4-RC1/log_processor.php
Revision: 201
Committed: Tue Jan 7 12:40:32 2014 UTC (10 years, 10 months ago) by matthys
Original Path: trunk/log_processor.php
File size: 11343 byte(s)
Log Message:
Rewrote bbc_update_host_stat($client); detect dns name <2, lenght dns name to display (last 2 or 3 parts)

File Contents

# User Rev Content
1 joku 63 <?php
2     /* This file is part of BBClone (A PHP based Web Counter on Steroids)
3     *
4     * SVN FILE $Id$
5     *
6     * Copyright (C) 2001-2014, the BBClone Team (see doc/authors.txt for details)
7     *
8     * This program is free software: you can redistribute it and/or modify
9     * it under the terms of the GNU General Public License as published by
10     * the Free Software Foundation, either version 3 of the License, or
11     * (at your option) any later version.
12     *
13     * This program is distributed in the hope that it will be useful,
14     * but WITHOUT ANY WARRANTY; without even the implied warranty of
15     * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16     * GNU General Public License for more details.
17     *
18     * See doc/copying.txt for details
19     */
20    
21     ///////////////////
22     // Log Processes //
23     ///////////////////
24    
25     // Checking where we came from
26     if (!defined("_MARK_PAGE")) return;
27    
28     // used by usort()
29     function bbc_sort_time_sc($row_a, $row_b) {
30     if ($row_a['time'] == $row_b['time']) return 0;
31     return ($row_a['time'] > $row_b['time'] ) ? 1 : -1;
32     }
33    
34     // return the key of a value
35     function bbc_get_key($array, $str) {
36     reset($array);
37    
38     while (list($idx, $val) = each($array)) {
39     if ($val == $str) return $idx;
40     }
41     return 0;
42     }
43    
44     // purge host and referrer stats at request
45     function bbc_purge_single() {
46     global $access;
47    
48     foreach (array("host", "key", "referer") as $cat) {
49     reset($access[$cat]);
50    
51     while (list($key, $score) = each($access[$cat])) {
52     if ($score == 1) {
53     unset($access[$cat][$key]);
54     ($cat == "referer") ? ++$access[$cat]['not_specified'] : "";
55     }
56     }
57     }
58     }
59    
60     // records the hosts that visited us most. Note, that we only pick up hostnames and strip
61     // any sort of prefix because else the listing would become rather useless
62     function bbc_update_host_stat($client) {
63     global $access, $BBC_IGNORE_BOTS;
64 matthys 201
65 joku 63 if ((empty($BBC_IGNORE_BOTS)) || (!isset($client['robot']))) {
66 matthys 201 $is_num = ($client['dns'] == $client['ip']) ? 1 : 0;
67     if ($is_num || strlen($client['dns']) < 2) {
68     // Numeric
69     if (strpos($client['ip'], ":") === false) {
70     // IPv4
71     $host = trim(substr($client['ip'], 0, strrpos($client['ip'], "."))).".&nbsp;-";
72     } else {
73     // IPv6
74     $parts = explode( ":", $client['ip']);
75     $parts = array_slice( $parts, 0, 4);
76     $host = implode( ":", $parts) . ":&nbsp;-";
77     }
78     } else {
79     // DNS name
80     $parts = explode( ".", $client['dns']);
81     $lastparts = implode( ".", array_slice( $parts, -2, 2) );
82     if (strlen($lastparts) <= 6) {
83     // Last 2 DNS name parts <= 6; take last 3 parts
84     $host = implode( ".", array_slice( $parts, -3, 3) );
85     } else {
86     // Last 2 DNS name parts > 6; take last 2 parts
87     $host = implode( ".", array_slice( $parts, -2, 2) );
88     }
89     }
90 joku 63 }
91     if ((empty($BBC_IGNORE_BOTS)) || (!isset($client['robot']))) {
92     if (!isset($access['host'][$host])) $access['host'][$host] = 0;
93     $access['host'][$host]++;
94     }
95    
96     if (isset($access['host']['not_specified'])) unset($access['host']['not_specified']);
97     }
98    
99     // the listing of the visited pages
100     function bbc_update_visits($time, $page, $nr) {
101     global $BBC_MAXVISIBLE, $last;
102    
103     $lv = count($last['traffic'][$nr]['views']) - 1;
104     $last_time = substr($last['traffic'][$nr]['views'][$lv], 0, strpos($last['traffic'][$nr]['views'][$lv], "|"));
105     $last_cnt = substr($last['traffic'][$nr]['views'][$lv], (strrpos($last['traffic'][$nr]['views'][$lv], "|") + 1));
106     $last_page = substr($last['traffic'][$nr]['views'][$lv], (strpos($last['traffic'][$nr]['views'][$lv], "|") + 1));
107     $last_page = substr($last_page, 0, strpos($last_page, "|"));
108     $last['traffic'][$nr]['off'] = !empty($last['traffic'][$nr]['off']) ? $last['traffic'][$nr]['off'] : 0;
109    
110     if ((empty($last['traffic'][$nr]['views'])) || (!is_array($last['traffic'][$nr]['views']))) return;
111    
112     if (intval($last_page) === intval($page)) {
113     $last['traffic'][$nr]['views'][$lv] = "$last_time|$last_page|".++$last_cnt;
114     $last['traffic'][$nr]['off']++;
115     }
116     else $last['traffic'][$nr]['views'][] = "$time|$page|1";
117    
118     sort($last['traffic'][$nr]['views']);
119    
120     // number of elements to be removed with array_splice() if necessary
121     $lv = count($last['traffic'][$nr]['views']) - 1;
122     $del = (($lv + 1) > $BBC_MAXVISIBLE) ? (($lv + 1) - $BBC_MAXVISIBLE) : false;
123     $last['traffic'][$nr]['views'] = ($del !== false) ? array_splice($last['traffic'][$nr]['views'], $del) :
124     $last['traffic'][$nr]['views'];
125     }
126    
127     // The most visited pages ranking
128     function bbc_update_page_stats($connect) {
129     global $access, $last;
130    
131     $long_page = $connect['page'];
132     $over_60 = (strlen($long_page) > 60) ? 1 : 0;
133     $connect['page'] = $over_60 ? "...".substr($long_page, -57) : $long_page;
134    
135     // Fix oversized page titles
136     if (($over_60) && (isset($access['page'][$long_page]['count']))) {
137     $access['page'][($connect['page'])]['count'] = $access['page'][$long_page]['count'];
138     $access['page'][($connect['page'])]['uri'] = $access['page'][$long_page]['uri'];
139     unset($access['page'][$long_page]);
140     }
141    
142    
143     if (!isset($access['page'][($connect['page'])]['count'])) {
144     $access['page'][($connect['page'])]['count'] = 0;
145     }
146    
147     $access['page'][($connect['page'])]['count']++;
148     $access['page'][($connect['page'])]['uri'] = $connect['uri'];
149    
150     $last['pages'] = ((empty($last['pages'])) || (!is_array($last['pages']))) ? array() : $last['pages'];
151    
152     if (($over_60) && (in_array($long_page, $last['pages']))) {
153     $last['pages'][bbc_get_key($last['pages'], $long_page)] = $connect['page'];
154     }
155     if (!in_array($connect['page'], $last['pages'])) $last['pages'][] = $connect['page'];
156    
157     $connect['page'] = bbc_get_key($last['pages'], $connect['page']);
158    
159     if (isset($connect['uri'])) unset($connect['uri']);
160    
161     return $connect;
162     }
163    
164     // Transfer the raw data from the main counters of var into $last.
165     // Any new data (more recent than $BBC_MAXTIME) is used in the global stats
166     function bbc_add_new_connections_to_old() {
167     global $BBC_IGNORE_AGENT, $BBC_IGNORE_BOTS, $BBC_MAXTIME, $BBC_MAXVISIBLE, $BBC_NO_DNS, $BBC_NO_HITS,
168     $BBC_PURGE_SINGLE, $access, $last;
169    
170     // Checking whether we have new connections
171     if (!$new_access = bbc_counter_to_array()) return false;
172    
173     // cleanup if requested
174     !empty($BBC_PURGE_SINGLE) ? bbc_purge_single() : "";
175    
176     ((!empty($access['time'])) && (is_array($access['time']))) ? bbc_time_offset() : "";
177    
178     // Upgrade from older versions. We need to erase the "last" data.
179     if (isset($access['last'])) unset($access['last']);
180    
181     //check for broken 0.4.2 referrer counting and apply fix if necessary
182     if (isset($access) && !isset($access['bugs']['ref_fix']) && isset($access['referer']['not_specified'])) {
183     bbc_fix_refer_stat(array_sum($access['referer']));
184     }
185    
186     // fix wrong browser assignments
187     foreach (array("java", "wwwc", "libwww") as $what) {
188     if (isset($access) && isset($access['stat']['browser'][$what])) {
189     $access['stat']['robot'][$what] = $access['stat']['browser'][$what];
190     $access['stat']['os']['other'] -= $access['stat']['robot'][$what];
191    
192     unset($access['stat']['browser'][$what]);
193     }
194     }
195    
196     $nb_new_access = (!empty($new_access) && is_array($new_access)) ? count($new_access) : 0;
197     $nb_last_access = (!empty($last['traffic']) && is_array($last['traffic'])) ? count($last['traffic']) : 0;
198    
199     foreach ($new_access as $connect) {
200     $connect = bbc_update_connect($connect);
201    
202     // the "last reset on" flag initialisation
203     if ((!isset($access['time'])) && (!isset($access['time']['reset']))) {
204     $access['time']['reset'] = $connect['time'];
205     }
206    
207     // Stop processing if bots are completely ignored
208     if ((!empty($BBC_IGNORE_BOTS)) && ($BBC_IGNORE_BOTS == 2)) {
209     if (!empty($connect['robot'])) {
210     --$nb_new_access;
211     continue;
212     }
213     }
214     // Omit referrers coming from robots
215     $connect['referer'] = !empty($connect['robot']) ? "unknown" : $connect['referer'];
216    
217     $this_connect = $connect['time'];
218     $last_connect = !empty($access['time']['last']) ? $access['time']['last'] : 0;
219    
220     // Hits as base for time stats if desired
221     if (empty($BBC_NO_HITS)) bbc_update_time_stat($this_connect, $last_connect);
222    
223     // The script viewed
224     $connect = isset($connect['page']) ? bbc_update_page_stats($connect) : $connect;
225     $prev_recorded = 0;
226    
227     // Check if a similar connection has been recorded yet
228     for($l = $nb_last_access - 1; ($l >= 0) && (($connect['time'] - $last['traffic'][$l]['time']) < $BBC_MAXTIME);
229     $l--) {
230     if (!empty($BBC_IGNORE_AGENT) ? ($connect['ip'] == $last['traffic'][$l]['ip']) :
231     (($connect['ip'] == $last['traffic'][$l]['ip']) && ($connect['agent'] == $last['traffic'][$l]['agent']))) {
232     $last['traffic'][$l]['page'] = $connect['page'];
233     $last['traffic'][$l]['time'] = $this_connect;
234     $last['traffic'][$l]['visits']++;
235     $access['stat']['totalvisits']++;
236    
237     ($BBC_MAXVISIBLE > 0) ? bbc_update_visits($connect['time'], $connect['page'], $l) : "";
238    
239     $prev_recorded = 1;
240     break;
241     }
242     }
243    
244     // Add new connection if it hasn't been recorded yet
245     if (!$prev_recorded) {
246     if (empty($access['stat']['totalvisits'])) $access['stat']['totalvisits'] = 0;
247     if (empty($access['stat']['totalcount'])) $access['stat']['totalcount'] = 0;
248    
249     $connect['dns'] = !empty($BBC_NO_DNS) ? $connect['ip'] : bbc_clean(gethostbyaddr($connect['ip']));
250     $connect['ext'] = bbc_get_extension($connect['dns'], $connect['ip']);
251    
252     $last['traffic'][$nb_last_access] = bbc_update_access($connect);
253     // Visit stats
254     $last['traffic'][$nb_last_access]['views'][] = $last['traffic'][$nb_last_access]['time']."|"
255     .$last['traffic'][$nb_last_access]['page']."|1";
256    
257     // Unique visits as base for time stats if desired
258     if (!empty($BBC_NO_HITS)) bbc_update_time_stat($this_connect, $last_connect);
259    
260     // Referrers collection will be updated all along with the keywords if available
261     if (isset($connect['referer'])) {
262     bbc_update_referer_stat($connect['referer']);
263     $flt_search = bbc_get_keywords($connect['referer']);
264     }
265     // The search as a whole in $last
266     $last['traffic'][$nb_last_access]['search'] = ($flt_search !== false) ? implode(" ", $flt_search) : "-";
267    
268     // The host listing
269     if ((isset($connect['dns'])) && (isset($connect['ip']))) {
270     bbc_update_host_stat($last['traffic'][$nb_last_access]);
271     }
272    
273     $access['stat']['totalvisits']++;
274     $access['stat']['totalcount']++;
275     $nb_last_access++;
276     }
277     }
278     return $nb_new_access;
279     }
280    
281     // Remove unnecessary connections from $last, that either exceed the $BBC_MAXVISIBLE limit or are
282     // older than time() - $BBC_MAXTIME.
283     function bbc_update_last_access() {
284     global $last, $BBC_MAXTIME, $BBC_MAXVISIBLE, $BBC_TIMESTAMP, $BBC_TIME_OFFSET;
285    
286     if (($BBC_MAXVISIBLE <= 0) || (empty($last['traffic'])) || (!is_array($last['traffic']))) {
287     $last['traffic'] = array();
288     return;
289     }
290     else {
291     $nb_connect = count($last['traffic']);
292     $ctime = $BBC_TIMESTAMP + ($BBC_TIME_OFFSET * 60);
293    
294     for ($k = $nb_connect - 1 - $BBC_MAXVISIBLE; $k >= 0; $k--) {
295     if (($ctime - $last['traffic'][$k]['time']) > $BBC_MAXTIME) unset($last['traffic'][$k]);
296     }
297     usort($last['traffic'],"bbc_sort_time_sc");
298     }
299     }
300 matthys 18 ?>

Properties

Name Value
svn:keywords Id